Title of the Talk: Application of machine learning to identify new potential drug candidates for Dravet syndrome
Dravet syndrome is an autosomal dominant genetic disorder which causes a catastrophic form of epilepsy. It is characterized by prolonged febrile and non-febrile seizures within the first year of life, which progress to other seizures, and psychomotor delay. In most cases, the mutations associated with the disease are not hereditary (nonsense mutations in the SCN1A gene, which codifies for NaV1.1). Accordingly, non selective anticonvulsant sodium channel blockers are known to aggravate seizures and status epilepticus in Dravet patients. Taking inspiration on recent success in the development of small molecules capable of selectively inhibiting the NaV1.7 subtype, we currently pursue the identification of highly selective NaV1.2 or NaV1.6 blockers that do not inhibit NaV1.1, as potential treatments for Dravet syndrome. For that purpose, machine learning models based on random subspace approximations, linear classifiers and ensemble learning, have been developed using NaV1.2 and NaV1.6 as targets and NaV1.1 as antitarget. The models have been applied in retrospective virtual screening campaigns, with very good enrichment metrics, and are currently applied complementary with structure-based approximations in the prospective virtual screening for new potential drug candidates against Dravet syndorme.
Prof. Alan Talevi was born in Buenos Aires, 1980. He obtained his Pharmacy degree in 2004 and completed his PhD studies in 2007, both at the University of La Plata (UNLP, Argentina). Among other recognitions, he obtained the award for the best PhD thesis on Computational Chemistry from the Argentinean Chemical Society (2008), a Pacifichem Young Scholar Award (USA, 2010) and the Award to Scientific and Technological Production from UNLP (2016). Since 2010, he holds a permanent position at the Argentinean Council of Scientific and Technical Research (CONICET), where he currently occupies an Independent Researcher position, and since 2012 he is the professor in charge of the Biopharmacy/Pharmacokinetics course at the Faculty of Exact Sciences, UNLP. He has published over 70 articles in periodicals and more than 30 book chapters, mostly in the fields of machine learning, drug discovery and biopharmacy. He was elected Head of the Department of Biological Sciences of the Faculty of Exact Sciences, UNLP, between 2015 and 2017. In 2018 he became the Head of the Laboratory of Bioactive Research and Development (LIDeB, UNLP). External reviewer for several agencies and universities, including the Argentinean National Institute of Cancer, Argentinean National Agency of Scientific and Technical Promotion, Argentinean Ministry of National Education, University of La República (Uruguay), University of Buenos Aires (Argentina), University of Córdoba (Argentina), University of Mar del Plata (Argentina), the National Agency of Scientific and Technical Promotion (Uruguay), the United Nations University, the National Science Center (Poland) and the “Victor Babes” University of Medicine and Pharmacy (Romania), CONCYTEC (Perú), and UK Research and Innovation (UK). Reviewer of more than 80 periodicals (https://publons.com/researcher/1667052/alan-talevi/), mostly in the fields of Cheminformatics, Drug Discovery, Drug Design and Pharmacology, including the Nature Communications, Scientific Reports, Journal of Medicinal Chemistry, Journal of Chemical Information and Modelling, Journal of Molecular Graphics and Modelling, PLoS One, Expert Opinion on Drug Delivery, Expert Opinion on Drug Discovery, Expert Opinion on Drug Metabolism and Toxicology, European Journal of Medicinal Chemistry, European Journal of Pharmacology, International Journal of Pharmaceutics, Journal of Pharmacy and Pharmacology. Outstanding reviewer of the European Journal of Medicinal Chemistry, European Journal of Pharmacology and Neuropharmacology, among others. Member of the editorial board of the Latin American Journal of Pharmacy (from 2014 and continues) and the Journal of Pharmacy and Nutrition Sciences. Guest editor of Mini-Reviews in Medicinal Chemistry, Frontiers in Chemistry and Current Drug Safety. Editor of the Antiepileptic Drug Discovery Novel Approaches volume of the Springer Protocols series (2016), and also of ADME Processes in Pharmaceutical Sciences Dosage, Design, and Pharmacotherapy Success (Springer, 2018). Chief Editor of the ADME Encyclopedia (Springer, 2021) and of the incoming Springer Nature Computer-Aided Drug Discovery and Design series.
Title of the Talk: Vibrotactile Coordinated Reset Stimulation for the Therapy of Parkinson’s Disease: Theory and Clinical Results
Abnormally strong synchronization of neuronal activity is a hallmark of several brain disorders, such as Parkinson’s disease (PD). In PD patients, permanent high-frequency deep brain stimulation suppresses symptoms during stimulus delivery. A qualitatively different, theory-based approach uses dedicated patterns of electrical stimuli delivered to brain tissue to cause a desynchronization-induced decoupling of oscillatory neurons, intending to move neural networks from attractors with abnormal neural synchrony to more physiological (desynchronized) attractors. Corresponding cumulative and long-lasting therapeutic and desynchronizing effects were demonstrated in animal and clinical studies.
By replacing invasively delivered electrical bursts with non-invasively administered vibratory bursts, we have developed a non-invasive approach: Vibrotactile Coordinated Reset (vCR) stimulation delivered to the fingertips was designed to treat PD patients by inducing long-lasting desynchronization. In a feasibility study, six PD patients were treated with vCR stimulation delivered for four hours per day for three months. Patients’ conditions were evaluated after medication withdrawal (‘off medication’) by means of standard clinical scores and EEG recordings before and after the 3-month vCR treatment. In accordance with vCR-related computational studies, months-long vCR therapy causes a statistically and clinically significant reduction of PD symptoms off medication together with a significant reduction of high beta (21-30 Hz) power in the sensorimotor cortex. The ultimate goal of this approach is to induce sustained symptom relief by non-invasively delivering desynchronizing stimulus patterns only regularly or occasionally.
Peter A. Tass studied medicine (MD, Ulm and Heidelberg Univ., Germany), physics (PhD, Stuttgart Univ., Germany), and mathematics (diploma, Stuttgart Univ.) and made a Habilitation in physiology (RWTH Aachen University, Germany). Since 2017 he is Professor of Neurosurgery in the Department of Neurosurgery, Stanford University Medical School, Stanford, CA, USA. Peter A. Tass investigates and develops neuromodulation techniques for understanding and treating neurologic conditions such as Parkinson’s disease, epilepsy, dysfunction following stroke and tinnitus. He creates invasive and non-invasive therapeutic procedures by means of comprehensive computational neuroscience studies and advanced data analysis techniques. Peter A. Tass has published more than 160 peer-reviewed papers, is inventor or co-inventor on more than 270 patents and received a number of national and international awards. Stanford lab homepage: http://med.stanford.edu/tass-lab.html
What to expect? Anticipation of contours and motion in early visual cortical circuits
The brain builds active predictions that allow inferring causal relationships in received sensory signals (Friston, 2005). These predictions are derived from the history of sensorimotor interactions with the world. Our perception can be understood as the result of real-time comparisons of information transmitted in sensory input signals with previously constructed activity patterns in the brain or priors (Sterzer et al., 2018). These patterns would explain why perceptual organization follows certain principles (Gestalt laws), as it happens during the conjunctions made between elements, necessary for the segmentation of surfaces or objects. The concept of priors is easily applicable to vision in two of its fundamental aspects: the perception of form and the perception of movement. Here, the possibility that simple visual priors are specifically constructed by the functional connectivity within the visual cortex will be discussed. In order to gather potential evidence for this hypothesis, data from our lab and others about the selectivity of long-range visual horizontal intrinsic and callosal connections, are presented and reviewed. Both networks selectively link distributed neuronal groups with similar response properties and the actions exerted by callosal input reflect the functional topography of those networks. Reversible deactivation studies strongly support that close to the representation of the visual field’s vertical meridian callosal networks affect in particular neurons which prefer potentially “midline crossing” features. Callosal connections seem to facilitate interactions between neurons with receptive fields along horizontal motion or (shape) trajectories crossing the visual field’s midline before a stimulus arrives. We propose that feature-selective lateral connections in general might exhibit a “cardinal bias” to interconnect neurons preferring vertical and horizontal contours. Those connections could support a spontaneously active network which pre-activates neurons along motion or shape trajectories frequently occurring in daily vision. Finally, the cardinal biases in frontal-eyed mammals are compared to the horizontal biases frequently observed in later-eyed mammals.
Currently Full Professor for Systems Neuroscience at the Brain Institute of the Federal University of Rio Grande do Norte (UFRN), Natal, Brazil. Director of the Brain Institute of the Federal University of Rio Grande do Norte (UFRN), Natal, Brazil. Scientific experience: Doctorate in medicine from Johann-Wolfgang Goethe Universität Frankfurt/Main and Max-Planck Institute (MPI) for Brain Research; Postdoc, MPI for Brain Research, Department for Neurophysiology; Research assistant, MPI for Brain Research, Department of Neurophysiology; Otto-Hahn fellowship at Karolinska Institute, Department of Neuroscience, Division of Neuroanatomy and Brain Development (Prof. Innocenti), Stockholm, Sweden; Independent Research Group Leader, MPI Brain Research, Frankfurt.
AutoML: The Machine Learning of the Machine Learning
With the increasing number of successful applications of Machine Learning algorithms in several application domains, including Computational Neuroscience, there is a strong interest in these algorithms easily accessible by users without Machine Learning expertise. There have been several efforts in this direction, involving not only the recommendation of the most suitable algorithm, but also their most appropriate hyper-parameter values and working pipeline. These efforts started a new research area, named Automated Machine Learning, AutoML, which has attracted the attention of researchers and practitioners not only from academia, but also from several companies working with Data Science and Machine Learning. This talk will present the main approaches and recent advances in this area, covering also works carried out in the Analytics Laboratory, at USP São Carlos.
Full Professor, Computer Science Department, University of São Paulo (USP), São Carlos, Brazil. Vice president of the Brazilian Computer Society. Prof. André de Carvalho has more than 300 publications in data science, machine learning and data mining, including 10 best papers awards from conferences organized by ACM, IEEE and SBC. Prof. André de Carvalho supervised/co-supervised more than 25 PhD students in different Universities in Brazil and Portugal and more than 15 Postdocs. He has been in the Editorial Board and Program Committee of some of the main conferences and Journals in Artificial Intelligence, Data Science, Data Mining, Machine Learning, like AAAI, KDD, ECML/PKDD, IJCAI and SDM. He is a member of the International Association for Statistical Computing (IASC) Council and Director of the Center of Machine Learning in Data Analysis, University of São Paulo.
Neuromodulatory Pattern Analysis for Local Filed Potentials
This talk will present a transient model to automatically detect neuromodulations in field potentials collected from the brain. The big advantage of the methodology is that it creates marked point processes directly from EEG, ECoG and LFPs that have high temporal resolution unlike the current approach using short term Fourier transforms. Moreover, the marked point processes for the neuromodulations can be easily integrated with neural firing providing multiscale quantification of brain activity, which facilitates relations to behavior. We will present several examples of the new model in behavior experiments.
Jose C. Principe (M’83-SM’90-F’00) is a Distinguished Professor of Electrical and Computer Engineering and Biomedical Engineering at the University of Florida where he teaches advanced signal processing, machine learning and artificial neural networks (ANNs) modeling. He is the Eckis Endowed Professor and the Founder and Director of the University of Florida Computational NeuroEngineering Laboratory (CNEL) www.cnel.ufl.edu . His primary area of interest is processing of time varying signals with adaptive neural models. The CNEL Lab has been studying signal and pattern recognition principles based on information theoretic criteria (entropy and mutual information)
Dr. Principe is an IEEE Fellow and received the prestigious IEEE Neural Network Pioneer Award in 2012. He was the past Chair of the Technical Committee on Neural Networks of the IEEE Signal Processing Society, Past-President of the International Neural Network Society, and Past-Editor in Chief of the IEEE Transactions on Biomedical Engineering. Dr. Principe has more than 900 publications, and an H-index of 82 (Google Scholar). He directed 98 Ph.D. dissertations and 65 Master theses. He wrote in 2000 an interactive electronic book entitled “Neural and Adaptive Systems” published by John Wiley and Sons and more recently co-authored several books on “Brain Machine Interface Engineering” Morgan and Claypool, “Information Theoretic Learning”, Springer, and “Kernel Adaptive Filtering”, Wiley.
The neurobiology of memory and forgetting
Memory is a biologically fundamental function essential for survival; it has the ability to retain information and recall it a later time. Furthermore, memories give us identity, guiding our thoughts and decisions, as well as influence our emotional behavior. Within the brain, memory is a dynamic property of populations of neurons and their interconnections. Our memories are not all created equally strong since some experiences are well remembered while others are not, depending on various factores such as emotion. Memories can be classified in explicit (e.g. declarative) and implicit (e.g. procedural). Memories are also classified according to time: from short-term memory, lasting only seconds or minutes, to long-term memory, lasting months or years, being that the last involves mechanisms more complex with the consolidation and recall. Each memory system has a distinct anatomical organization, where different parts of the brain are recruited during phases of memory storage. Is this normal memory loss or the onset of dementia? Normal forgetting, in contrast to pathological forgetting that occurs in disease and with aging, is necessary for the functioning of a healthy brain as important as the ability to remember. The importance of forgetting, which involves the loss of information over time, is a relatively new concept for science. From numerous fields revealing that there are separate mechanisms biochemistry and molecular within neurons and, probably, also within glia (astrocytes): some for memory and the other for forgetting. These findings point to new way that helps us to undertand the neurobiology of memorry and forgetting.
CNPq Research Productivity Scholarship - Level-1A. Full Professor of Biochemistry, ICBS, Federal University of Rio Grande do Sul (UFRGS), MSc and Ph.D. in UFRGS. Full Member of the Brazilian Academy of Sciences (ABC) in Biological Sciences. Full Member of the World Academy of Sciences (TWAS) in Medicine and Health Sciences (Neurosciences). She works in the areas of Biochemistry and Neuroscience, with an emphasis on neuroprotection and hereditary neurometabolic diseases. Angela has published more than 400 scientific articles in reference international journals in her field and formed more than 70 masters and doctors. She participated in the scientific committee of funding bodies such as CNPq and FAPERGS and was part of the editorial board of scientific journals Metabolic Brain Disease, Molecular Neurobiology, and Neurotoxicity Research. She has been Regional Secretary of SBPC-RS since 2019. She was awarded the CAPES-ELSEVIER Award, which recognizes researchers and their contribution to the development of scientific researchers who in their academic career present a production of high impact and excellence portrayed in the database Scopus, was awarded the Gaucho Researcher Award - FAPERGS 2018 and recently received "The Scientist of the Year 2020 Award" from the International Achievements Research Center, Chicago, USA. For more than 5 years, she has been developing a project for the dissemination of science, focusing on neuroscience and interdisciplinarity, for children from Public Schools in Porto Alegre, with the aim of encouraging a taste for study, science, and health.